solarwindpy 0.0.1.dev0__py3-none-any.whl → 0.1.1__py3-none-any.whl

plans/tests-audit/artifacts/ARCHITECTURE_RECOMMENDATIONS.md

@@ -0,0 +1,943 @@
+# Architecture Test Recommendations - SolarWindPy Enhancement
+
+## Executive Summary
+
+Based on the Architecture Compliance Analysis (Grade: B+), this document provides **actionable recommendations for 42 specific architecture tests** that will enhance DataFrame structure validation, memory efficiency, and performance benchmarking in the SolarWindPy test suite.
+
+**Target Coverage Increase**: +5-6% (from 77.1% to 82-83%)  
+**Estimated Implementation**: 8-12 hours  
+**Priority**: High (foundational architecture validation)
+
+---
+
+## 1. MultiIndex Structure Validation Tests (12 tests)
+
+### 1.1 Core Structure Tests (4 tests)
+
+**File**: `tests/core/test_multiindex_validation.py` (NEW)
+
+#### Test 1: `test_multiindex_column_names_required`
+```python
+def test_multiindex_column_names_required():
+    """Test that MultiIndex columns must have correct names."""
+    # Test with missing names
+    tuples = [("v", "x", "p1"), ("b", "y", "")]
+    bad_index = pd.MultiIndex.from_tuples(tuples)  # No names
+    
+    # Should warn or fail when creating DataFrame with unnamed levels
+    with pytest.warns(UserWarning, match="MultiIndex levels should be named"):
+        # Test DataFrame creation validation
+        pass
+```
+
+#### Test 2: `test_multiindex_level_order_validation`  
+```python
+def test_multiindex_level_order_validation():
+    """Test that MultiIndex levels must be in correct order."""
+    # Test incorrect level order (S, C, M instead of M, C, S)
+    tuples = [("p1", "x", "v"), ("", "y", "b")]
+    wrong_order = pd.MultiIndex.from_tuples(tuples, names=["S", "C", "M"])
+    
+    # Should detect incorrect level ordering
+    with pytest.raises(ValueError, match="MultiIndex levels must be in order"):
+        # Test validation function
+        pass
+```
+
+#### Test 3: `test_multiindex_empty_component_scalar_validation`
+```python
+def test_multiindex_empty_component_scalar_validation():
+    """Test proper handling of empty components for scalar measurements."""
+    # Density (scalar) should have empty component
+    density_tuples = [("n", "", "p1"), ("n", "", "p2")]
+    
+    # Velocity (vector) should not have empty component
+    with pytest.raises(ValueError, match="Vector measurements require components"):
+        velocity_tuples = [("v", "", "p1")]  # Invalid: missing component
+```
+
+#### Test 4: `test_multiindex_empty_species_bfield_validation`
+```python
+def test_multiindex_empty_species_bfield_validation():
+    """Test proper handling of empty species for magnetic field."""
+    # Magnetic field should have empty species
+    bfield_tuples = [("b", "x", ""), ("b", "y", ""), ("b", "z", "")]
+    
+    # Ion measurements should not have empty species
+    with pytest.raises(ValueError, match="Ion measurements require species"):
+        ion_tuples = [("n", "", "")]  # Invalid: missing species
+```
+
+### 1.2 Hierarchy Consistency Tests (4 tests)
+
+#### Test 5: `test_measurement_component_consistency`
+```python
+def test_measurement_component_consistency():
+    """Test M-C level consistency rules."""
+    # Scalar measurements (n, w scalar) must have empty C
+    # Vector measurements (v, b) must have x,y,z components
+    # Tensor measurements (w tensor) must have par,per components
+    
+    valid_combinations = [
+        ("n", "", "p1"),      # Valid: scalar density
+        ("v", "x", "p1"),     # Valid: vector velocity
+        ("w", "par", "p1"),   # Valid: tensor thermal speed
+        ("b", "z", ""),       # Valid: vector magnetic field
+    ]
+    
+    invalid_combinations = [
+        ("n", "x", "p1"),     # Invalid: scalar with component
+        ("v", "", "p1"),      # Invalid: vector without component
+        ("b", "par", ""),     # Invalid: wrong component for B-field
+    ]
+```
+
+#### Test 6: `test_species_measurement_consistency`
+```python
+def test_species_measurement_consistency():
+    """Test M-S level consistency rules."""
+    # Magnetic field (b) must have empty species
+    # Ion measurements must have valid species
+    # Multi-species combinations validation
+    
+    valid_species = ["p1", "p2", "a", "e"]
+    invalid_species = ["", "x", "invalid"]
+```
+
+#### Test 7: `test_multiindex_dataframe_creation_validation`
+```python
+def test_multiindex_dataframe_creation_validation():
+    """Test DataFrame creation with various MultiIndex configurations."""
+    # Test successful creation with proper structure
+    # Test failure modes with invalid structures
+    # Test warning generation for questionable structures
+```
+
+#### Test 8: `test_multiindex_level_access_patterns`
+```python
+def test_multiindex_level_access_patterns():
+    """Test proper level access methods."""
+    # Test level="M" vs level=0 (should prefer level name)
+    # Test xs() vs iloc/loc patterns
+    # Test IndexSlice usage patterns
+```
+
+### 1.3 Integration Tests (4 tests)
+
+#### Test 9: `test_plasma_multiindex_integration`
+```python
+def test_plasma_multiindex_integration():
+    """Test Plasma class MultiIndex requirements."""
+    # Test Plasma object creation with various MultiIndex configurations
+    # Test validation of input DataFrames
+    # Test error handling for invalid structures
+```
+
+#### Test 10: `test_ion_multiindex_integration`
+```python
+def test_ion_multiindex_integration():
+    """Test Ion class MultiIndex requirements."""
+    # Test Ion object creation with species-specific data
+    # Test validation of M-C level consistency for ions
+    # Test error handling for invalid ion data structures
+```
+
+#### Test 11: `test_spacecraft_multiindex_integration`
+```python
+def test_spacecraft_multiindex_integration():
+    """Test spacecraft data MultiIndex patterns."""
+    # Test coordinate system MultiIndex structures
+    # Test position/velocity MultiIndex validation
+    # Test spacecraft data integration with plasma data
+```
+
+#### Test 12: `test_cross_module_multiindex_consistency`
+```python
+def test_cross_module_multiindex_consistency():
+    """Test MultiIndex consistency across modules."""
+    # Test data exchange between Plasma, Ion, Spacecraft
+    # Test MultiIndex preservation during operations
+    # Test consistency validation across module boundaries
+```
+
+---
+
+## 2. Memory Efficiency Tests (10 tests)
+
+### 2.1 View vs Copy Tests (4 tests)
+
+**File**: `tests/core/test_dataframe_memory.py` (NEW)
+
+#### Test 13: `test_xs_returns_view_not_copy`
+```python
+def test_xs_returns_view_not_copy():
+    """Test that xs() operations return views, not copies."""
+    data = create_test_plasma_data()
+    
+    # Test that xs() returns view
+    view = data.xs("v", level="M")
+    original_value = data.iloc[0, 0]
+    
+    # Modify view
+    view.iloc[0, 0] = original_value + 1000
+    
+    # Original should be modified (proving it's a view)
+    assert data.iloc[0, 0] == original_value + 1000
+```
+
+#### Test 14: `test_memory_usage_comparison_view_vs_copy`
+```python
+def test_memory_usage_comparison_view_vs_copy():
+    """Compare memory usage of view vs copy operations."""
+    large_data = create_large_test_data(n_points=50000)
+    
+    # Measure view memory
+    view = large_data.xs("v", level="M")
+    view_memory = view.memory_usage(deep=True).sum()
+    
+    # Measure copy memory  
+    copy = large_data.xs("v", level="M").copy()
+    copy_memory = copy.memory_usage(deep=True).sum()
+    
+    # View should use significantly less memory
+    assert view_memory < copy_memory * 0.1  # View should be <10% of copy
+```
+
+#### Test 15: `test_chained_xs_operations_memory`
+```python
+def test_chained_xs_operations_memory():
+    """Test memory efficiency of chained xs() operations."""
+    data = create_test_plasma_data()
+    
+    # Single operation
+    single_op = data.xs("v", level="M").xs("p1", level="S")
+    
+    # Chained operations should not create intermediate copies
+    memory_before = get_process_memory()
+    result = data.xs("v", level="M").xs("p1", level="S")
+    memory_after = get_process_memory()
+    
+    memory_increase = memory_after - memory_before
+    assert memory_increase < 1.0  # Less than 1MB increase
+```
+
+#### Test 16: `test_dataframe_slice_memory_efficiency`
+```python
+def test_dataframe_slice_memory_efficiency():
+    """Test memory efficiency of different slicing methods."""
+    data = create_test_plasma_data()
+    
+    # Compare memory usage of different access patterns
+    patterns = [
+        lambda df: df.xs("v", level="M"),
+        lambda df: df.loc[:, pd.IndexSlice["v", :, :]],
+        lambda df: df[df.columns[df.columns.get_level_values("M") == "v"]],
+    ]
+    
+    # Test that xs() is most memory efficient
+```
+
+### 2.2 Large Dataset Tests (3 tests)
+
+#### Test 17: `test_large_multiindex_dataframe_memory`
+```python
+def test_large_multiindex_dataframe_memory():
+    """Test memory usage with large MultiIndex DataFrames."""
+    sizes = [1000, 10000, 100000]
+    memory_usages = []
+    
+    for size in sizes:
+        data = create_test_data(n_points=size)
+        memory = data.memory_usage(deep=True).sum()
+        memory_usages.append(memory)
+    
+    # Memory should scale linearly with data size
+    assert_linear_scaling(sizes, memory_usages, tolerance=0.2)
+```
+
+#### Test 18: `test_memory_usage_by_data_type`
+```python
+def test_memory_usage_by_data_type():
+    """Test memory usage with different data types."""
+    dtypes = [np.float32, np.float64, np.int32, np.int64]
+    
+    for dtype in dtypes:
+        data = create_test_data(dtype=dtype)
+        memory = data.memory_usage(deep=True).sum()
+        
+        # Verify expected memory usage based on dtype
+        expected_memory = calculate_expected_memory(data.shape, dtype)
+        assert abs(memory - expected_memory) / expected_memory < 0.1
+```
+
+#### Test 19: `test_memory_cleanup_after_operations`
+```python
+def test_memory_cleanup_after_operations():
+    """Test that temporary DataFrames are properly cleaned up."""
+    initial_memory = get_process_memory()
+    
+    # Perform operations that create temporary objects
+    for i in range(100):
+        data = create_test_data()
+        result = data.xs("v", level="M").xs("p1", level="S")
+        processed = result.groupby(level=0).mean()
+        del data, result, processed
+    
+    gc.collect()
+    final_memory = get_process_memory()
+    
+    # Memory should return close to initial level
+    memory_increase = final_memory - initial_memory
+    assert memory_increase < 10.0  # Less than 10MB permanent increase
+```
+
+### 2.3 Memory Profiling Tests (3 tests)
+
+#### Test 20: `test_multiindex_operation_memory_profile`
+```python
+@memory_profiler.profile
+def test_multiindex_operation_memory_profile():
+    """Profile memory usage of common MultiIndex operations."""
+    data = create_large_test_data()
+    
+    # Profile common operations
+    operations = [
+        lambda: data.xs("v", level="M"),
+        lambda: data.loc[:, pd.IndexSlice["v", :, "p1"]],
+        lambda: data.groupby(level=["M", "S"]).mean(),
+        lambda: data.reindex(columns=data.columns.sort_values()),
+    ]
+    
+    for op in operations:
+        memory_before = get_process_memory()
+        result = op()
+        memory_after = get_process_memory()
+        
+        # Log memory usage for each operation
+        memory_delta = memory_after - memory_before
+        assert memory_delta < 100.0  # Reasonable memory usage
+```
+
+#### Test 21: `test_memory_leak_detection`
+```python
+def test_memory_leak_detection():
+    """Test for memory leaks in DataFrame operations."""
+    baseline_memory = get_process_memory()
+    
+    # Repeat operations many times
+    for iteration in range(1000):
+        data = create_test_data()
+        
+        # Perform various operations
+        v_data = data.xs("v", level="M")
+        species_data = v_data.xs("p1", level="S")
+        stats = species_data.describe()
+        
+        # Clean up explicitly
+        del data, v_data, species_data, stats
+        
+        # Check memory every 100 iterations
+        if iteration % 100 == 0:
+            gc.collect()
+            current_memory = get_process_memory()
+            memory_growth = current_memory - baseline_memory
+            
+            # Should not have significant memory growth
+            assert memory_growth < 50.0  # Less than 50MB growth
+```
+
+#### Test 22: `test_dataframe_copy_detection`
+```python
+def test_dataframe_copy_detection():
+    """Test detection of unexpected DataFrame copies."""
+    data = create_test_data()
+    original_id = id(data._mgr)  # Access internal data manager
+    
+    # Operations that should return views
+    view_ops = [
+        lambda: data.xs("v", level="M"),
+        lambda: data.loc[:10],
+        lambda: data.iloc[:10],
+    ]
+    
+    for op in view_ops:
+        result = op()
+        # Verify result shares data with original (implementation dependent)
+        # This test helps identify when operations unexpectedly create copies
+```
+
+---
+
+## 3. Performance Benchmark Tests (8 tests)
+
+### 3.1 Access Pattern Benchmarks (3 tests)
+
+**File**: `tests/core/test_dataframe_performance.py` (NEW)
+
+#### Test 23: `test_multiindex_access_performance`
+```python
+def test_multiindex_access_performance():
+    """Benchmark different MultiIndex access patterns."""
+    data = create_large_test_data(n_points=100000)
+    
+    # Test access methods
+    access_methods = {
+        'xs_method': lambda: data.xs("v", level="M"),
+        'loc_indexslice': lambda: data.loc[:, pd.IndexSlice["v", :, :]],
+        'boolean_mask': lambda: data[data.columns.get_level_values("M") == "v"],
+    }
+    
+    # Benchmark each method
+    results = {}
+    for name, method in access_methods.items():
+        times = []
+        for _ in range(10):
+            start = time.time()
+            result = method()
+            end = time.time()
+            times.append(end - start)
+        
+        results[name] = np.mean(times)
+    
+    # xs() should be fastest
+    assert results['xs_method'] < results['boolean_mask']
+    assert results['xs_method'] < results['loc_indexslice']
+```
+
+#### Test 24: `test_hierarchical_groupby_performance`
+```python
+def test_hierarchical_groupby_performance():
+    """Test performance of groupby operations on MultiIndex."""
+    data = create_large_test_data(n_points=100000)
+    
+    groupby_operations = [
+        lambda: data.groupby(level="M").mean(),
+        lambda: data.groupby(level=["M", "S"]).mean(),  
+        lambda: data.groupby(level=["M", "C", "S"]).mean(),
+    ]
+    
+    # Test that performance scales reasonably with groupby complexity
+    for i, op in enumerate(groupby_operations):
+        start = time.time()
+        result = op()
+        duration = time.time() - start
+        
+        # More complex groupbys should take longer but reasonably so
+        assert duration < 10.0  # Should complete within 10 seconds
+```
+
+#### Test 25: `test_multiindex_sorting_performance`
+```python
+def test_multiindex_sorting_performance():
+    """Test performance of MultiIndex sorting operations."""
+    # Create unsorted data
+    data = create_test_data(sorted=False)
+    
+    sorting_operations = [
+        lambda: data.sort_index(axis=1),
+        lambda: data.sort_index(axis=1, level="M"),
+        lambda: data.sort_index(axis=1, level=["M", "C", "S"]),
+    ]
+    
+    # Test sorting performance
+    for op in sorting_operations:
+        start = time.time()
+        result = op()
+        duration = time.time() - start
+        
+        assert duration < 5.0  # Should complete within 5 seconds
+        assert result.columns.is_monotonic_increasing
+```
+
+### 3.2 Scalability Tests (3 tests)
+
+#### Test 26: `test_dataframe_size_scalability`
+```python
+def test_dataframe_size_scalability():
+    """Test DataFrame operation performance at different scales."""
+    sizes = [1000, 10000, 100000]
+    operation_times = []
+    
+    for size in sizes:
+        data = create_test_data(n_points=size)
+        
+        start = time.time()
+        # Standard operations
+        v_data = data.xs("v", level="M")
+        p1_data = v_data.xs("p1", level="S")
+        stats = p1_data.describe()
+        end = time.time()
+        
+        operation_times.append(end - start)
+    
+    # Performance should scale sub-quadratically
+    assert_subquadratic_scaling(sizes, operation_times)
+```
+
+#### Test 27: `test_multiindex_column_count_scalability`
+```python
+def test_multiindex_column_count_scalability():
+    """Test performance with increasing column counts."""
+    column_counts = [10, 50, 100, 500]
+    access_times = []
+    
+    for count in column_counts:
+        data = create_test_data(n_columns=count)
+        
+        start = time.time()
+        result = data.xs("v", level="M")
+        duration = time.time() - start
+        
+        access_times.append(duration)
+    
+    # Access time should scale linearly with column count
+    assert_linear_scaling(column_counts, access_times, tolerance=0.3)
+```
+
+#### Test 28: `test_datetime_index_scalability`
+```python
+def test_datetime_index_scalability():
+    """Test performance with large datetime indices."""
+    durations = ['1D', '7D', '30D', '365D']  # Data spanning different periods
+    
+    for duration in durations:
+        data = create_time_series_data(duration=duration, freq='1min')
+        
+        start = time.time()
+        # Common time series operations
+        hourly = data.resample('1H').mean()
+        daily = data.resample('1D').mean() 
+        end = time.time()
+        
+        operation_time = end - start
+        assert operation_time < 30.0  # Should complete within 30 seconds
+```
+
+### 3.3 Regression Tests (2 tests)
+
+#### Test 29: `test_performance_regression_baseline`
+```python
+def test_performance_regression_baseline():
+    """Establish performance baselines for regression testing."""
+    # Standard test data
+    data = create_standard_test_data()
+    
+    # Baseline operations with expected completion times
+    baselines = {
+        'xs_access': (lambda: data.xs("v", level="M"), 0.01),  # 10ms
+        'species_select': (lambda: data.xs("p1", level="S"), 0.01),  # 10ms  
+        'component_select': (lambda: data.xs("x", level="C"), 0.01),  # 10ms
+        'groupby_mean': (lambda: data.groupby(level="M").mean(), 0.1),  # 100ms
+    }
+    
+    for name, (operation, max_time) in baselines.items():
+        times = []
+        for _ in range(5):
+            start = time.time()
+            result = operation()
+            end = time.time()
+            times.append(end - start)
+        
+        avg_time = np.mean(times)
+        assert avg_time < max_time, f"{name} took {avg_time:.3f}s, expected <{max_time}s"
+```
+
+#### Test 30: `test_memory_usage_regression_baseline`
+```python
+def test_memory_usage_regression_baseline():
+    """Establish memory usage baselines for regression testing."""
+    # Standard operations with expected memory usage
+    data = create_standard_test_data()
+    
+    memory_baselines = {
+        'full_dataframe': (lambda: data, 10.0),  # 10MB
+        'velocity_data': (lambda: data.xs("v", level="M"), 2.0),  # 2MB
+        'single_species': (lambda: data.xs("p1", level="S"), 5.0),  # 5MB
+    }
+    
+    for name, (operation, max_mb) in memory_baselines.items():
+        result = operation()
+        memory_mb = result.memory_usage(deep=True).sum() / 1024**2
+        
+        assert memory_mb < max_mb, f"{name} used {memory_mb:.1f}MB, expected <{max_mb}MB"
+```
+
+---
+
+## 4. Edge Case & Integration Tests (7 tests)
+
+### 4.1 Index Consistency Tests (3 tests)
+
+**File**: `tests/core/test_dataframe_edge_cases.py` (NEW)
+
+#### Test 31: `test_datetime_index_validation`
+```python
+def test_datetime_index_validation():
+    """Test DateTime index validation and naming."""
+    # Test proper datetime index
+    good_index = pd.date_range('2020-01-01', periods=100, freq='1min')
+    good_index.name = 'Epoch'
+    data = create_test_data(index=good_index)
+    
+    # Test invalid indices
+    bad_indices = [
+        pd.RangeIndex(100),  # Non-datetime
+        pd.date_range('2020-01-01', periods=100)[::-1],  # Non-monotonic
+        pd.date_range('2020-01-01', periods=100, name='time'),  # Wrong name
+    ]
+    
+    for bad_index in bad_indices:
+        with pytest.warns(UserWarning):
+            invalid_data = create_test_data(index=bad_index)
+```
+
+#### Test 32: `test_multiindex_duplicate_handling`
+```python
+def test_multiindex_duplicate_handling():
+    """Test handling of duplicate MultiIndex columns."""
+    # Create data with duplicate columns
+    tuples = [("v", "x", "p1"), ("v", "x", "p1"), ("b", "y", "")]
+    duplicate_columns = pd.MultiIndex.from_tuples(tuples, names=["M", "C", "S"])
+    
+    # Test detection and handling of duplicates
+    with pytest.raises(ValueError, match="Duplicate columns detected"):
+        data = pd.DataFrame(np.random.randn(100, 3), columns=duplicate_columns)
+        validate_multiindex_structure(data)
+```
+
+#### Test 33: `test_missing_level_handling`
+```python
+def test_missing_level_handling():
+    """Test handling of missing MultiIndex levels."""
+    # Test missing M level
+    tuples = [("", "x", "p1"), ("", "y", "p1")]
+    missing_m = pd.MultiIndex.from_tuples(tuples, names=["M", "C", "S"])
+    
+    # Test missing C level  
+    tuples = [("v", "", "p1"), ("v", "", "p2")]
+    missing_c = pd.MultiIndex.from_tuples(tuples, names=["M", "C", "S"])
+    
+    # Test validation of missing required levels
+    for invalid_index in [missing_m, missing_c]:
+        with pytest.raises(ValueError, match="Required level values missing"):
+            validate_multiindex_structure(pd.DataFrame(columns=invalid_index))
+```
+
+### 4.2 Data Type Edge Cases (2 tests)
+
+#### Test 34: `test_mixed_data_types_multiindex`
+```python
+def test_mixed_data_types_multiindex():
+    """Test MultiIndex with mixed data types."""
+    # Create DataFrame with mixed types
+    data = pd.DataFrame({
+        ('n', '', 'p1'): np.random.randn(100).astype(np.float32),
+        ('v', 'x', 'p1'): np.random.randn(100).astype(np.float64),
+        ('w', 'par', 'p1'): np.random.randn(100).astype(np.int32),
+    })
+    data.columns = pd.MultiIndex.from_tuples(data.columns, names=["M", "C", "S"])
+    
+    # Test that operations work correctly with mixed types
+    result = data.xs("p1", level="S")
+    assert isinstance(result, pd.DataFrame)
+    
+    # Test type preservation
+    assert data[('n', '', 'p1')].dtype == np.float32
+    assert data[('v', 'x', 'p1')].dtype == np.float64
+```
+
+#### Test 35: `test_nan_handling_multiindex_operations`
+```python
+def test_nan_handling_multiindex_operations():
+    """Test NaN handling in MultiIndex operations."""
+    data = create_test_data()
+    
+    # Introduce NaN values
+    data.iloc[10:20, :] = np.nan
+    data.iloc[:, 2] = np.nan  # Entire column
+    
+    # Test that MultiIndex operations handle NaN correctly
+    v_data = data.xs("v", level="M")
+    assert v_data.isna().any().any()  # Should contain NaN
+    
+    # Test statistics with NaN
+    stats = v_data.describe()
+    assert not stats.isna().all().any()  # Should compute valid statistics
+    
+    # Test groupby with NaN
+    grouped = data.groupby(level="M").mean()
+    assert isinstance(grouped, pd.DataFrame)
+```
+
+### 4.3 Integration Edge Cases (2 tests)
+
+#### Test 36: `test_cross_module_dataframe_compatibility`
+```python
+def test_cross_module_dataframe_compatibility():
+    """Test DataFrame compatibility across modules."""
+    # Create data in plasma format
+    plasma_data = create_test_plasma_data()
+    
+    # Test Ion object creation
+    ion_data = plasma_data.xs("p1", level="S")
+    ion = ions.Ion(ion_data, "p1")
+    assert isinstance(ion, ions.Ion)
+    
+    # Test Spacecraft data integration
+    sc_data = create_test_spacecraft_data()
+    combined = pd.concat([plasma_data, sc_data], axis=1)
+    
+    # Test that combined data maintains MultiIndex structure
+    assert combined.columns.nlevels == 3
+    assert list(combined.columns.names) == ["M", "C", "S"]
+```
+
+#### Test 37: `test_large_dataset_edge_cases`
+```python
+def test_large_dataset_edge_cases():
+    """Test edge cases with large datasets."""
+    # Create very large dataset
+    large_data = create_test_data(n_points=1000000)  # 1M points
+    
+    # Test memory-efficient operations
+    subset = large_data.iloc[::1000]  # Sample every 1000th point
+    v_subset = subset.xs("v", level="M")
+    
+    # Test that operations complete without memory errors
+    stats = v_subset.describe()
+    assert isinstance(stats, pd.DataFrame)
+    
+    # Test chunked operations
+    chunk_size = 10000
+    chunk_stats = []
+    for i in range(0, len(large_data), chunk_size):
+        chunk = large_data.iloc[i:i+chunk_size]
+        chunk_stat = chunk.xs("v", level="M").mean()
+        chunk_stats.append(chunk_stat)
+    
+    # Verify chunked processing works
+    assert len(chunk_stats) > 1
+```
+
+---
+
+## 5. Specialized Architecture Tests (5 tests)
+
+### 5.1 IndexSlice Pattern Tests (2 tests)
+
+**File**: `tests/core/test_indexslice_patterns.py` (NEW)
+
+#### Test 38: `test_indexslice_multiindex_selection`
+```python
+def test_indexslice_multiindex_selection():
+    """Test IndexSlice patterns for MultiIndex selection."""
+    data = create_test_data()
+    
+    # Test various IndexSlice patterns
+    patterns = [
+        pd.IndexSlice["v", :, :],           # All velocity data
+        pd.IndexSlice["v", "x", :],         # All x-components of velocity
+        pd.IndexSlice["v", :, "p1"],        # All velocity components for p1
+        pd.IndexSlice[["v", "b"], :, :],    # Velocity and magnetic field
+        pd.IndexSlice[:, ["x", "y"], :],    # x and y components only
+    ]
+    
+    for pattern in patterns:
+        result = data.loc[:, pattern]
+        assert isinstance(result, pd.DataFrame)
+        
+        # Verify selection worked correctly
+        if pattern == pd.IndexSlice["v", :, :]:
+            assert all(result.columns.get_level_values("M") == "v")
+```
+
+#### Test 39: `test_complex_indexslice_combinations`
+```python
+def test_complex_indexslice_combinations():
+    """Test complex IndexSlice combinations."""
+    data = create_test_data()
+    
+    # Complex selection patterns
+    complex_patterns = [
+        # Velocity x,y components for protons
+        (pd.IndexSlice["v", ["x", "y"], ["p1", "p2"]], 
+         lambda df: (df.columns.get_level_values("M") == "v") & 
+                   df.columns.get_level_values("C").isin(["x", "y"]) &
+                   df.columns.get_level_values("S").isin(["p1", "p2"])),
+        
+        # All vector components (exclude scalars)
+        (pd.IndexSlice[:, ["x", "y", "z"], :],
+         lambda df: df.columns.get_level_values("C").isin(["x", "y", "z"])),
+    ]
+    
+    for pattern, validator in complex_patterns:
+        result = data.loc[:, pattern]
+        expected_mask = validator(data)
+        expected = data.loc[:, expected_mask]
+        
+        pd.testing.assert_frame_equal(result.sort_index(axis=1), 
+                                    expected.sort_index(axis=1))
+```
+
+### 5.2 Performance Optimization Tests (3 tests)
+
+#### Test 40: `test_multiindex_operation_optimization`
+```python
+def test_multiindex_operation_optimization():
+    """Test optimization of common MultiIndex operations."""
+    data = create_large_test_data()
+    
+    # Compare optimized vs unoptimized approaches
+    approaches = {
+        'optimized_xs': lambda: data.xs("v", level="M").xs("p1", level="S"),
+        'chained_loc': lambda: data.loc[:, pd.IndexSlice["v", :, "p1"]],
+        'boolean_filter': lambda: data[
+            (data.columns.get_level_values("M") == "v") & 
+            (data.columns.get_level_values("S") == "p1")
+        ],
+    }
+    
+    # Time each approach
+    times = {}
+    for name, approach in approaches.items():
+        start = time.time()
+        result = approach()
+        times[name] = time.time() - start
+    
+    # Optimized approach should be fastest
+    assert times['optimized_xs'] <= min(times.values()) * 1.1  # Within 10%
+```
+
+#### Test 41: `test_dataframe_reindexing_performance`
+```python
+def test_dataframe_reindexing_performance():
+    """Test performance of DataFrame reindexing operations."""
+    data = create_test_data(sorted=False)
+    
+    # Test different reindexing strategies
+    strategies = [
+        lambda df: df.sort_index(axis=1),
+        lambda df: df.reindex(columns=df.columns.sort_values()),
+        lambda df: df[df.columns.sort_values()],
+    ]
+    
+    results = []
+    for strategy in strategies:
+        start = time.time()
+        result = strategy(data)
+        duration = time.time() - start
+        results.append((duration, result))
+        
+        # Verify result is correctly sorted
+        assert result.columns.is_monotonic_increasing
+    
+    # All strategies should complete in reasonable time
+    for duration, _ in results:
+        assert duration < 1.0  # Less than 1 second
+```
+
+#### Test 42: `test_memory_efficient_aggregations`
+```python
+def test_memory_efficient_aggregations():
+    """Test memory-efficient aggregation operations."""
+    data = create_large_test_data()
+    
+    # Test aggregations that should be memory efficient
+    aggregations = [
+        ('mean_by_measurement', lambda df: df.groupby(level="M").mean()),
+        ('std_by_species', lambda df: df.groupby(level="S").std()),
+        ('sum_by_component', lambda df: df.groupby(level="C").sum()),
+    ]
+    
+    for name, aggregation in aggregations:
+        memory_before = get_process_memory()
+        result = aggregation(data)
+        memory_after = get_process_memory()
+        
+        memory_increase = memory_after - memory_before
+        data_memory = data.memory_usage(deep=True).sum() / 1024**2
+        
+        # Memory increase should be reasonable compared to data size
+        assert memory_increase < data_memory * 2  # Less than 2x data size
+        assert isinstance(result, pd.DataFrame)
+```
+
+---
+
+## 6. Implementation Plan
+
+### 6.1 Phase 1: Core Infrastructure (Tests 1-12, 31-33)
+**Duration**: 3-4 hours  
+**Files**: 
+- `tests/core/test_multiindex_validation.py`
+- `tests/core/test_dataframe_edge_cases.py`
+
+**Priority**: High - foundational validation
+
+### 6.2 Phase 2: Memory & Performance (Tests 13-30)
+**Duration**: 4-5 hours  
+**Files**:
+- `tests/core/test_dataframe_memory.py`  
+- `tests/core/test_dataframe_performance.py`
+
+**Priority**: High - performance bottleneck identification
+
+### 6.3 Phase 3: Advanced Patterns (Tests 34-42)
+**Duration**: 2-3 hours
+**Files**:
+- `tests/core/test_indexslice_patterns.py`
+- Additional tests in existing files
+
+**Priority**: Medium - optimization and edge cases
+
+### 6.4 Integration & Validation
+**Duration**: 1 hour
+- Run full test suite
+- Validate coverage increase
+- Performance regression testing
+
+---
+
+## 7. Expected Outcomes
+
+### 7.1 Coverage Metrics
+- **Current**: 77.1% overall coverage
+- **Target**: 82-83% overall coverage  
+- **New Architecture Tests**: 42 tests
+- **Estimated Coverage Increase**: +5-6%
+
+### 7.2 Quality Improvements
+- **Architecture Validation**: Comprehensive MultiIndex structure testing
+- **Memory Efficiency**: Systematic memory usage validation
+- **Performance Baselines**: Regression testing infrastructure
+- **Edge Case Coverage**: Robust handling of boundary conditions
+
+### 7.3 Maintenance Benefits
+- **Regression Prevention**: Performance and memory baselines
+- **Development Velocity**: Clear architecture compliance validation
+- **Documentation**: Living examples of proper DataFrame usage
+- **Onboarding**: Clear patterns for new developers
+
+---
+
+## 8. Metrics & Success Criteria
+
+### 8.1 Quantitative Metrics
+- [ ] 42 new architecture tests implemented
+- [ ] Test suite coverage ≥82%
+- [ ] Architecture compliance grade improves to A-
+- [ ] Memory usage tests coverage ≥90%
+- [ ] Performance regression tests established
+
+### 8.2 Qualitative Metrics  
+- [ ] Clear MultiIndex validation patterns documented
+- [ ] Memory efficiency best practices established
+- [ ] Performance baselines for regression testing
+- [ ] Edge case handling comprehensively tested
+- [ ] Integration patterns validated across modules
+
+---
+
+*Generated: 2025-08-21*  
+*DataFrameArchitect Agent - Architecture Enhancement Recommendations*  
+*SolarWindPy Test Suite Audit - Phase 3 Deliverable*